Failure Mode of Fastener... 1

[mrbebu]

Hello all,

Experienced fastener failures recently and wanted to get some input as to what could've been the cause. Before I give more context and detail, I would just like to see what the trained eyes of this forum have to says solely in regards to the pattern.


Pics are of the same fastener, different angles:

 

[azcats]

Prefaced with I'm no expert, but...

That looks like it was loaded in a manner that put the bolt in cyclical bending inducing fatigue, beaching and ultimately a fracture.

What's the loading like? Fasteners tensioned? Slip between the faying surfaces?

Got ahead of myself - I'll stay tuned for your update.

 

ASM Handbook Volume 11: Failure Analysis and Prevention

This handbook has a handy visual chart that links fatigue fracture appearance resulting from various modes of loading and mean stresses.

This will just get you to the fracture mechanism (obviously fatigue), which does not explain why it failed. I recommend a metallurgical lab for that.


"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[Ron]

azcats is correct. Cyclic loading evidenced by the clean beach marks on the right followed by fast failure on the left. The clean beach mark area indicates low stress, high cycles to failure in my opinion.

 

[desertfox]

I agree with a fatigue failure, however checking the bolt preload on assembly might avoid such a failure in future.
Any idea what the bolts are tightened too? also how many bolts are in the joint? increasing bolt numbers and or size might also avoid such a failure.

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[mrbebu]

Thank you everyone for your input.

Now to give more detail:

This is an automotive wheel joint that was in a endurance circuit racing application (think 24 Hours of LeMans but on an amateur level with converted street cars). The studs were recently installed with approximately 9 total hours of track time prior to the failure.

It is a typical 5 bolt joint with "high performance" M12 x 1.5 sized fasteners heat treated to a Rc hardness of 38-40 (185+ ksi tensile). The threads are rolled, not cut, but I'm unsure yet if they are rolled pre or post heat treat, which I've read makes a significant difference in fatigue resistance.

They also have a Molybdenum DiSulfide dry lubricant coating on them and I was told that they were torqued to 88 lbs-ft prior to racing.

The joint members are the steel hub (~22-25 Rc) and the wheel; forged/machined aluminum alloy 6061-T6 with a standard 60 degree tapered lug nut seat. Lug nut seats have no paint or finish. No steel seat inserts.
----------------------------------------------------------------------------------------

So, it seems that the consensus for root cause of the failure is insufficient preload/installation torque. It's obviously difficult to know for sure what the end user really torqued them to, but understanding how variable the torque tightening method for desired preload is, it's the easiest thing to point to.

My question is, how do I determine that it actually truly was that and not a defect in the fasteners that the cracks initiated from? I suppose sending the failed studs to a metallurgist is my only option?

 

You've now introduced human safety as a potential direct consequence of final failure, which for a wheel stud tends only to happen at high load, i.e., high speed or in a corner.

I urge you to retain a qualified metallurgical laboratory to perform an investigation. Eng-tips has some very qualified contributors, but failure analysis is not done by crowdsourcing.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[kingnero]

So, it seems that the consensus for root cause of the failure is insufficient preload/installation torque.

I fail to see how you got that conclusion from this thread.

88 lbs-ft seems plenty, if not on the high side.
Improper fastening of a nut on a wheel will more likely result in loosening of the nut (and shortly afterwards, of the remainder of the fasteners).

x2 on the advice given above.

 

[Tmoose]

"9 total hours of track time prior to the failure."

Was the racing done at 9 mph or less ?

"Remember to re-torque the nuts or bolts after 40 kilometers (25 miles) of use. Failure to re-torque the nut or bolts is unsafe and could cause a serious accident."

"CAUTION TO OWNERS OF MAG WHEELS, ALLOY WHEELS OR CUSTOM WHEELS: The wheels lug nuts must be retorqued within 100 miles and checked periodically or an extremely hazardous condition may result. Flynn’s Tire will retorque these lug nuts within 100 miles at no charge and assumes no responsibility for any damage or injury that may occur if the customer fails to have the lug nuts checked as recommended."

Hot Rod golf cart wheels -

https://fairwayalloys.com/pages/customer-warranty-care

"On all installations, proper torque specifications must be used. You must return to the installer after 10 miles and check/retorque your wheels if necessary. Failure to do so is extremely unsafe and could cause an accident resulting in serious bodily injury and/or property damage."

=============

Were the lug nuts the proper 60° taper, or spherical ? Did they include a separate rotating element ?
I would insist on seeing all the hubs, all the studs, all the nuts and all the wheels.
I would be armed with bright lights, magnification, and several cans of the chemicals used for liquid penetrant inspection.
I fully expect there is some critical evidence visible on those components.

Taper seats directly on aluminum is a crime against nature in my opinion.
Spherical/ball seats are better

 

[desertfox]

Hi mrbebu

The conclusion that it was a fatigue failure was reached by looking at the fracture surfaces and studying well known fracture pictures of other fatigue like failures, however incorrect preload is one possible answer to the cause of failure, it might be that the bolts had a manufacturing flaw that we don’t know about.
I would heed Ironic’s advice and get those bolts to a lab, from the information you have given the cause of incorrect preload cannot be proved we would need the external loads that the joint would see during service.

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[mrbebu]

You've now introduced human safety as a potential direct consequence of final failure, which for a wheel stud tends only to happen at high load, i.e., high speed or in a corner.

I urge you to retain a qualified metallurgical laboratory to perform an investigation. Eng-tips has some very qualified contributors, but failure analysis is not done by crowdsourcing.

I fail to see how you got that conclusion from this thread.

My mistake, didn't mean to come across stating that insufficient preload/installation torque as being the definite. It's at/near the top of the list of probable causes based on much of the literature I'm reviewing and persons I'm conversing with.

I will most certainly send the failed specimens to a metallurgical lab. Can anyone recommend a lab on the east coast of the US?

The highest loaded corner on a car in a turn is typically the outside front. Notably the left front on race tracks in North America (predominantly clockwise direction). The kicker is wheel stud failures I know of/seen almost always happen on the rear axle of rear wheel driven cars (note I said "failures" as this is not an isolated incident. This occurrence is actually common across differing cars/teams during these endurance racing events). My intuition as to why the rear has a higher frequency is because of constant reversed lateral loading from engine torque and braking over the course of an endurance race.

Were the lug nuts the proper 60° taper, or spherical ? Did they include a separate rotating element ?
I would insist on seeing all the hubs, all the studs, all the nuts and all the wheels.
I would be armed with bright lights, magnification, and several cans of the chemicals used for liquid penetrant inspection.
I fully expect there is some critical evidence visible on those components.

Taper seats directly on aluminum is a crime against nature in my opinion.
Spherical/ball seats are better

Unfortunately, I don't have that accessibility. The car and parts are located on the opposite coast. The vast majority of aluminum wheels (aftermarket and OEM) have a 60° taper seats without steel inserts, at least on cars.

No on the separate rotating element lug nuts. No one uses those in racing. But, plastic deformation of the aluminum under the lug nut causing loss of preload is a big concern as you point out. The wheels were new from a quality manufacturer, but not brand new for the race. I'm told they had two test days with being removed/installed and retorqued on a couple of times.

As for checking torque on new wheels after some usage, there's a stigma for torquing/checking lug nut torque during pit stops when the car comes in hot from a racing stint. The idea is that the studs are "hot" and can potentially be overstretched if you put torque on them. The temperature at the wheel/studs/hub interface isn't much more than 200F under normal operation, which shouldn't have a significant affect on the Youngs Modulus of the stud alloy (4340), unless I'm missing something, which can very much be the case. Otherwise there's a larger underlying issue generating excessive heat.

I argue that risking having insufficient preload by not checking torque is much worse than worrying about overstretching from torquing while hot. I just tell people to lower their torque settings ~15% to be safe during pitstops, but admittedly that value is just a guess (to be clear, this end user didn't check torque throughout the race). The big unknown is how friction is affected by the dry lubricant coating at elevated temperatures. A reputable fastener company mentioned that at temperatures above 100F, these coatings can have a significant reduction in friction.

My plan in the meantime is to try and establish a torque value for a desired preload by measuring stud stretch in the lab with a wheel hub and a center section of a wheel cutout. I will also do this at elevated temperature levels to see what affect it has. I know the torque/preload scatter will most likely be high, but it's better than nothing since I get various torque recommendations from fastener manufacturers that supply wheel studs.

If there are any other suggestions for testing, please, lay it on me thick. I have access to a well equipped tool room and machine shop.

 

[kingnero]

My (amateur) experience on the track says that the front hub gets well over 100°C. Are you talking about oval racing, where the brakes are less solicited than at conventional tracks?

 

I will most certainly send the failed specimens to a metallurgical lab. Can anyone recommend a lab on the east coast of the US?

I don't know any names, but your problem is a going concern in the trucking industry. You might want to start by seeking out engineering consultants specializing in vehicle safety and accident reconstruction. They won't generally have labs, but will work with them. Your state DOT almost certainly has guidelines for this and may be able to steer (sorry) you right.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[desertfox]

Hi mrbebu

Unless you know the loading in terms of force on the wheel in service you cannot predict what the bolt preload should actually be, that said I assume somewhere that an analysis of the wheels on a race car and its loading have been done and it’s that information you need,

Now you mention in a more recent post about the wheels being tested and stripped down several times and re-torqued , so did they use new or the existing nuts when the wheel was reassembled? Reason for asking is if an existing nut/bolt is used then a figure for the preload would or should be below 70% the yield stress, if new bolts were used then that would mean the bolts were originally tightened to give a preload figure of 90% yield stress and should be replaced.
In a nutshell I am saying that if existing bolts have been re-used when in fact they should have been replaced, then this fact alone may also contribute to the failure mode.

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[Tmoose]

https://res.cloudinary.com/engtips/..._faces:center/https://i.imgur.com/Vbb0CDY.jpg

It looks to me like there are multiple crack initiations at the 1;30 to 4;30 o'clock position on the stud.
Is that the horse pressure driving torque edge, or the brake torque edge?
If the wheels are hub centric, then I'd kind of expect that cornering forces would mostly tug axially on the studs.
Whereas braking/accelerating would try to bend loose studs, an cause crax to initiate as shown in the picture.

What is the thickness of the wheel thru which the stud passes?
What is the detail on the hub side of the wheel around each stud hole ? A Simple small chamfer at each stud hole ? A full flat face ? Sculptured pockets leaving a "pad" to contact the hub?

https://encrypted-tbn0.gstatic.com/...fpUlFFmbFwg0Lh7p_-4NKplXJVdmUx1shCEq2YhocN_&s

https://images.moderntiredealer.com/post/M-OFFSET-Mounting-plate-1-2.jpg

How thick is the brake rotor?

How thick is the hub in the vicinity of the studs?
Is the hub sculptured on the backside, so the hub is thin in between the stud holes?

 

[mrbebu]

My (amateur) experience on the track says that the front hub gets well over 100°C. Are you talking about oval racing, where the brakes are less solicited than at conventional tracks?

Not oval, "road" racing as we like to call it in the states. 2500-2800lbs cars with 300+bhp. Shooting the wheel studs by the hub with an infrared temp sensor after a stint I was only seeing 220-240F max. I should disclose that running 2 piece style brake rotors with aluminum centers helps tremendously with this. Wheel bearings won't survive very long much higher than that temperature, AFAIK.

I don't know any names, but your problem is a going concern in the trucking industry. You might want to start by seeking out engineering consultants specializing in vehicle safety and accident reconstruction. They won't generally have labs, but will work with them. Your state DOT almost certainly has guidelines for this and may be able to steer (sorry) you right

I've contacted a metallurgist testing facility that my "day job" has used in years past. They quoted me $3500 for a failure analysis report. A bit steep for a vanity project, but at the very least I will be sending pieces for material strength and chemical composition analysis.

Hi mrbebu

Unless you know the loading in terms of force on the wheel in service you cannot predict what the bolt preload should actually be, that said I assume somewhere that an analysis of the wheels on a race car and its loading have been done and it’s that information you need,

Now you mention in a more recent post about the wheels being tested and stripped down several times and re-torqued , so did they use new or the existing nuts when the wheel was reassembled? Reason for asking is if an existing nut/bolt is used then a figure for the preload would or should be below 70% the yield stress, if new bolts were used then that would mean the bolts were originally tightened to give a preload figure of 90% yield stress and should be replaced.
In a nutshell I am saying that if existing bolts have been re-used when in fact they should have been replaced, then this fact alone may also contribute to the failure mode.

I think it is safe to assume that wheel studs and nuts are most certainly treated as reused fasteners so a max of 75% of proof load.

https://res.cloudinary.com/engtips/image/fetch/w_8...

It looks to me like there are multiple crack initiations at the 1;30 to 4;30 o'clock position on the stud.
Is that the horse pressure driving torque edge, or the brake torque edge?
If the wheels are hub centric, then I'd kind of expect that cornering forces would mostly tug axially on the studs.
Whereas braking/accelerating would try to bend loose studs, an cause crax to initiate as shown in the picture.

What is the thickness of the wheel thru which the stud passes?
What is the detail on the hub side of the wheel around each stud hole ? A Simple small chamfer at each stud hole ? A full flat face ? Sculptured pockets leaving a "pad" to contact the hub?

https://encrypted-tbn0.gstatic.com/images?q=tbn:AN...

https://images.moderntiredealer.com/post/M-OFFSET-...

How thick is the brake rotor?

How thick is the hub in the vicinity of the studs?
Is the hub sculptured on the backside, so the hub is thin in between the stud holes?

Looks to be horsepower driving edge with crack formation and propagation in that direction (hub rotates counter clockwise from engine torque w.r.t. to the picture below). Missing stud was also broken, but lost in the frenzy during the race:

The constant reversed torque cycles acting perpendicular to the axis of the stud is why I think the rear has been more prone to potential loosening and then failure.

Correct on hubcentric wheel and lateral/cornering forces pulling axially on the studs. The friction interface should ideally be resisting engine torque and braking torque.

Wheel thickness is 16-17mm from the hub face to the top of the 60 degree cone where the lug nut sits.

Hub side of wheel face:

Brake rotor thickness is 7-8mm.

The hub is uniform thickness with small pockets on the backside to allow the stud head to sit parallel to the hub face. Hub is 0.375" (9.5mm) thick where studs install.

-------------------------------------------------------------------------------------------


So in the meantime I've started to check bolt stretch w.r.t. torque settings (10 lbs-ft increments) on a bench setup. I'm using 3 different lug nuts, 2 have a dry lubricant coating and are very popular in racing and the other 3 are just a typical chrome like finish with no lubricants. The particular wheel is an OEM aluminum wheel with a 60° taper seat with no steel/s.steel inserts.

I'm discovering the comment earlier about taper seats directly on aluminum being a crime has some real validity (no surprise). Especially with dry lug nuts. I'm finding that anything passed a particular torque value has no effect on bolt stretch, just a large increase in friction (I'm also measuring loosening torque).

I'm sourcing another aluminum wheel that is very popular in the track/racing world as I suspect the wheel I'm using, although OEM, is softer than what is typically used on track (need to check hardness on it still). I will be repeating the tests with lubrication on the lug seats and threads, as well as at elevated temperatures.

 

[3DDave]

Since it's a bending load, that suggests the rim is shifting in rotation, causing a moment in the stud. The load is concentrated by the lug nut. It look like there is a nice blend at the hub which avoids a concentration there.

You may have reached a point where the clamping friction available between the rim and the hub is not enough to keep them from moving relative to each other. This might be aggravated if the aluminum is plastically deforming and lowering the pre-load and some amount from thermal expansion/contraction.

 

[BridgeSmith]

You may have reached a point where the clamping friction available between the rim and the hub is not enough to keep them from moving relative to each other.

This seems to be a likely scenario. You may have to consider adding some type of radial shear keys or shear pins at the mating surface to lock the wheel from rotating relative to the hub.

Rod Smith, P.E., The artist formerly known as HotRod10

 

[Tmoose]

Got pictures of both sides of the rotor and the actual wheel mounting face?

It looks like the hub spline drags up/upsets material above the hub surface around each stud.
Is the brake rotor shiny or scuffed around each stud hole, like it was shuffling in service and shined up the raised lips on the hub face?

That is starting to look like the very definition of "embedment," if any of the 4 faying surfaces of the hub, brake rotor, and wheel start off with line contact.
And just a few .001s inches of embedment can easily cause a huge loss of preload/clamp force even from scientifically tightened fssteners. With resulting bending loads no fastener should be asked to endure.

A chamfer on the hub side of the rotor might help short term.

A larger chamfer on the hub's stud holes might keep the spline from upsetting hub material higher than the rotor mounting face.

 

[mrbebu]

I really think this may be an embedment related incident as well considering what I'm observing on my bench testing rig with the aluminum wheel lug seat.

Stud clearance holes on rotors are typically very large, like +2mm or so since it relies on the hub for locating. I don't have the rotor that was on the car with me. If you zoom in on the picture of the hub, you can see a clean area around the studs/holes, but the original machining marks are still present and not abraded. I think it


So fastener experts,question about determining the effective length for elongation measurements. Since the lug nut seats are 60 degree taper, how would I handle the necessary parameters (notably "Lj") in the effective length formula per Machinerys' Handbook.

Here's a modeled cross section of the joint:

Here's the referenced formula:

Thoughts?